Loyalty beyond points: what three retailers actually built

🇹🇷 Türkçe: Bu yazının Türkçesini oku →

A retention conversation with a mid-market brand almost always starts in the same place. How many points to a redemption. What the cash value of a point is. Whether the birthday freebie still moves anything. That is a fine conversation, but it is not the one the retailers who are actually winning at retention are having with themselves right now.

The brands that are pulling ahead stopped thinking of loyalty as a discounting mechanism years ago. They treat it as the operating layer for their first-party customer data, and everything visible to the shopper — the tier, the challenge, the birthday gift, the in-app game — is a user interface on top of that layer. Points sit inside it as one lever among many. They are not the point.

I want to walk through three brands who put enough hard numbers in the public record in the last twelve months that you can see the shape of the machine, not just the marketing. Sephora, Starbucks, Chipotle. Then we will get to what a mid-market store on WordPress and WooCommerce can copy this quarter, without the six-figure loyalty platform that all three of them can afford and you probably cannot.

Sephora’s 45 million members, and the interesting number is 6%

Sephora reported 45 million Beauty Insider members in North America at the end of 2025, growing to close to 46 million shortly after. Emmy Berlind, Sephora’s Senior Vice President and General Manager of Loyalty, is on the record with the shape of the pyramid: Rouge, the top tier, is 6% of the base. Everything else — Insider at the bottom, VIB in the middle — is the other 94%. That 6% is where the money is, and the retention work is entirely about earning the top-tier shopper’s next visit rather than pulling a new one off the street with a discount.

Two things Sephora did in 2025 are worth studying. The first was a Rouge Celebration event, run without a sitewide discount. Berlind, speaking to BeautyMatter in October 2025, framed it as “developing the emotional connection” and reported that average order value rose 40% versus the prior year during the event, with double-digit year-over-year sales growth. The second was Beauty Insider Challenges, a gamified layer of missions and rewards that had reached 30% of the U.S. member base within roughly a year of launch — a data point Forbes carried in its 9 February 2026 report.

• 45 million Beauty Insider members reported in the January 2026 BusinessWire release.
• Rouge tier: 6% of the member base, drives disproportionate revenue.
• 40% AOV lift on the Rouge Celebration event, no sitewide discount attached (Berlind, BeautyMatter, October 2025).
• 30% of members playing Beauty Insider Challenges (Forbes, February 2026).

The transferable lesson is unglamorous. Sephora spent its 2025 loyalty budget on a small number of top-tier shoppers and on gamified engagement that costs almost nothing to deliver but produces a lot of first-party data about what an individual member likes. The lever that was not pulled — a broad discount — is the one every board asks for.

Starbucks brought tiers back because points on their own had stopped moving frequency

Starbucks Rewards had 34.6 million 90-day active U.S. members reported in the Q1 fiscal 2025 filing published 28 January 2025, and the same programme reached 35.5 million by Q1 fiscal 2026. In fiscal year 2025 Rewards members drove nearly 60% of U.S. company-operated revenue, which the company put at more than 13 billion dollars of member spend. The programme is one of the more mature loyalty machines in retail.

And Starbucks still decided, in early 2026, that a flat single-tier programme was not enough. At Investor Day on 29 January 2026 the company announced a three-tier structure — Green, Gold, Reserve — that launched on 10 March 2026. Tressie Lieberman, Starbucks Global Chief Brand Officer, called it in the official announcement “a key milestone in our Back to Starbucks strategy” that would “reinvigorate what it means to be a Starbucks Rewards member.”

Read the mechanics carefully and it is not primarily a points change. The 60-star redemption for two dollars off is generous, but the load-bearing pieces are non-expiring Stars for Gold and Reserve members as long as they hold status, escalating earning rates that reward frequency rather than basket size, and a Reserve level unlocked at 2,500 stars a year with curated events and trips. The company is buying frequency and identity, and using the tier ladder to segment its own data.

• 35.5 million U.S. active members at Q1 FY26; nearly 60% of U.S. company-operated revenue in FY25, more than $13 billion.
• Three-tier structure (Green, Gold, Reserve) launched 10 March 2026 as part of a wider “Back to Starbucks” reset.
• Non-expiring stars at Gold and Reserve; earning rates rise from one to 1.7 stars per dollar across the tiers.

The transferable lesson is that when a single-tier programme has drifted into “everyone gets the same thing”, the right response is not more generous points. It is a re-segmentation of the base that makes it worth someone’s while to move up a level. Points are the currency; the tier is the product.

Chipotle rebuilt Rewards to close the gap between the app and the counter

Chipotle had a good problem and a bad problem at the same time. The good one: roughly 90% of Chipotle app transactions link to a Rewards account. The bad one: only about 20% of in-restaurant transactions do. Curt Garner, Chipotle’s President and Chief Strategy and Technology Officer, has been open about the gap in every 2025 earnings call. The company relaunched Rewards on 13 April 2026 as “Rewards on Repeat”, published as an official newsroom release with 21 million active members named in the announcement.

The mechanics are worth quoting because they are cleaner than most relaunches I have read this year. New members get complimentary chips and guacamole on a five-dollar-plus first purchase. Points now expire only after a full year without a qualifying purchase, up from six months. Birthday rewards get a 30-day redemption window, not a same-day burn. Freepotle bonus drops return monthly. Redemption thresholds came down across the board. In Garner’s words, “we’re delivering more rewards to all members, more often, and enhancing existing benefits.”

• 21 million active Chipotle Rewards members named in the April 2026 relaunch release; grew to 23 million by mid-2026.
• Roughly 90% of digital transactions link to a Rewards account; roughly 20% of in-restaurant transactions do.
• Points expire only after a year without a qualifying purchase; birthday reward window extended to 30 days.

The transferable lesson is that the loyalty programme’s job is to close whichever data gap is costing you the most. Chipotle is not primarily solving for “how do I reward my best customers”; it is solving for “how do I get the in-store visit into the same data set as the app order”. Everything in the relaunch — the extended point life, the friendlier birthday window, the monthly Freepotle drop — exists to make it worth the shopper’s while to scan on a counter visit rather than an app one. The reward is the mechanism; the data is the objective.

What the three have in common is not the points

Three programmes, three different mechanics, one shared shape. In each case the loyalty programme is a data acquisition and retention layer, and the visible rewards are the price of the data. Sephora is buying deep engagement from a small, high-value segment through experiences and gamified challenges. Starbucks is buying frequency and identity through a tier ladder that costs it very little to operate. Chipotle is buying in-store identification of shoppers it already knows online.

Two things are absent from all three stories, and they are the two things buyers ask us about first. The first is a headline sitewide discount as the primary lever. Sephora put its 2025 Rouge event on record precisely because it moved a 40% AOV lift without one. The second is a heroic technology purchase framed as the answer. All three of these programmes sit on top of platforms the company already had; the work was in the model of the customer, not the vendor slide.

If we were briefing a mid-market brand on a loyalty relaunch tomorrow, the honest sequence looks like this. Name the specific data gap the programme is supposed to close (Sephora: engagement from the top 6%; Starbucks: frequency across tiers; Chipotle: identification in-store). Design the reward layer to make it worth the shopper’s while to close that gap. Publish an internal metric that is legible to the CFO — AOV lift, share of revenue from members, share of transactions identified — and hold the programme to it. Points are downstream of all three of those decisions.

What this looks like inside a real WooCommerce store

Here is the build we would put in front of a mid-market brand on WordPress and WooCommerce, without buying a new platform. Five items, each of which we have shipped on real stores, in an order that mirrors the priority above.

1. Extend the customer record beyond email and order history. The first-party attribute that pays back fastest is not the one your ESP asks for; it is the one you can only get by asking the shopper directly during account creation or a post-purchase flow — first purchase category, birthday month, size, use case. Two or three attributes, captured cleanly, piped from user meta to the ESP as segments that survive a re-platforming. Budget: a fortnight if the data model is already clean, longer if the account area has drifted.

2. Build a real customer journey surface in the account area. The default WooCommerce “My Account” page is transactional. Replace it with a page that shows the shopper their own story: order history grouped by category, saved products, tier progress if you go that route, referral history, next redemption. This is the first thing a top-tier shopper ever sees, and it costs a mid-sized brand about ten to fifteen thousand euros to do well. Sephora’s account page is not magic; the difference is that it treats the shopper as someone with a history rather than a session.

3. A challenge or mission does not need a loyalty platform to run. A Beauty Insider Challenges equivalent — buy from three sub-categories in a month for a bonus, review two purchases, try a new brand — is WooCommerce meta plus Action Scheduler plus a small UI. The work is not the engineering; it is the merchandising discipline to design challenges that produce useful data rather than gamified giveaways. Budget it as a merchandising ticket, not an IT one.

4. Publish a single retention metric to leadership monthly. Pick one of the three shapes named above — share of revenue from identified members, AOV differential between members and non-members, or share of transactions identified — and put it in the same monthly board pack the CFO already reads. Every brand we have seen make real progress on retention has this line. Every brand stuck at points-and-freebies is measuring loyalty in isolation from the P&L.

5. Tie the programme to margin, not revenue. This is the discipline that keeps a discount-heavy programme from eating you alive. Every loyalty spend — points issued, birthday freebie cost, event cost — gets attributed against retained margin from the segment it targets, not against topline. Sephora runs its Rouge tier this way. Starbucks runs its Reserve level this way. A mid-market brand can run it this way on WooCommerce with a monthly reconciliation job; the discipline matters more than the tooling.

Effort estimate for the five together: one focused quarter with a WooCommerce team of two developers, a merchandiser and an analyst. Cost order of magnitude: five figures in build, not six; the ongoing operating cost lives in the merchandising team, not the platform bill.

If we were writing the brief for a first meeting with a mid-market brand on this, we would put three questions on the table. What specific data gap is your loyalty programme designed to close, and can you name it in one sentence. What is the metric on the CFO’s monthly pack that this programme is measured against. If the programme were switched off tomorrow, what proportion of your retained margin would you lose. None of those questions is about the number of points to a redemption; and that is the point.